Problems with a purpose
Implement the pieces you actually use: broadcasting, gradient descent, normalization, attention, and more.
YOUR NEXT REP STARTS HEREYou know the theory.
You know the theory.
Now write the code.
From your first tensor to a transformer from scratch. Build real ML intuition, one working implementation at a time.
1,868 practice problems Real PyTorch Free during beta
scaled_dot_product.pyPYTORCH
MediumScaled dot-product attention
Turn the equation into something that runs.
def attention(q, k, v):
# All you need is a few good primitives.
d = q.shape[-1]
scores = q @ k.transpose(-2, -1)
scores = scores / math.sqrt(d)
weights = torch.softmax(scores, dim=-1)
return weights @ vA peek inside your practice workspace
MAKE IT SECOND NATURE
Tensors Autograd Neural networks Attention Optimization
LESS WATCHING. MORE BUILDING.The missing reps between
The missing reps between
understanding and doing.
Reading a paper and implementing it are different skills. This is where you practice the second one.
Feedback you can act on
Run sample checks as you work. Submit against the full test suite and see exactly where your next attempt should focus.
Build a habit that compounds
Follow a study plan, save a tricky problem, and pick up your code where you left it. Your progress stays with your account.
A PATH FOR EVERY STAGE
All study plans Start where you are.
Make one concept
your own today.
Find your first problem Browse freely. Create an account to run code and save your progress.